This article was originally reviewed for my statistics and research methods course.
Kellsey, C. & Knievel, J. (2012). Overlap between humanities faculty citation and library monograph collections, 2004-2009. College & Research Libraries, 73(6), 569-583.
The purpose of this study was to determine whether library collections are meeting the needs of faculty members, as assessed by citation analysis of books published by authors in four humanities departments. The researchers were also interested in investigating several related research questions: How old are the books that scholars were using? How were those books acquired by the libraries (via approval or firm order)? How interdisciplinary are the sources that scholars are examining?
The researchers in this study were very thorough in their literature review – they found previous studies that touched on all of their research questions, and they were able to integrate this information into an examination of the field that didn’t feel like a list of names and results. I also found their method of citation analysis intriguing. Although I know that citation analysis is something often done in the “harder” sciences, I felt that they adapted the method well for humanities data. They were also careful to mention the limitations of what their method could account for, particularly in their explanation of unavailable data in regard to dates of faculty activity or book purchases. I disagreed with them that this data would have been particularly difficult to find – there are records kept of when people are employed and books are bought, even before computers – but I also think that these missing pieces come out in the wash, so to speak. Overall, their method and their data seems relatively sound.
This study did have some weaknesses, but most of them were acknowledged by the authors, who encourage other researchers to replicate their study in order to build the pool of available data. They mention that their averages can be skewed by outlying data, which caused me to question why they didn’t use a more precise statistical method (although it’s possible that my own lack of familiarity with such methods may be blinding me to the fact that they used the most effective method they could)?
I was also at times confused at the choices that they made in inclusion or exclusion of their sample books. On page 573 the authors list the books that they excluded, which include multiple edited volumes; on the next page they state that “Chapters or articles in edited compiled volumes were counted as books” (Kellsey and Knievel, p. 574) for purposes of citation analysis. Why were edited volumes not used as books in the sample populations, but citations of similar books are considered books during the analysis phase? In addition, it was unclear to me why chapters or articles in books wouldn’t have been considered articles, as opposed to books in their own rights.
This article was clear in the implications and applications of this research for libraries, specifically for those librarians concerned with collection development. The authors pointed out that the usual standards for weeding and moving to storage might not be helpful in keeping resources available for faculty, and that libraries might recreate the study with their own collections and re-examine those workflows in light of this new data.
I do have to say that, while I found this study to be well-thought-out and executed, the language was frequently ambiguous and at times it was difficult to decipher the authors’ meaning. Their first research question, “…not whether anyone used what we already own, but instead whether we own what was needed” (Kellsey and Knievel, p. 569) seems a bit like trying to prove a negative (logically impossible) and could have been framed in a way that was more clear. They also never actually specify at what institution they are doing this research, nor do they define terms that they mention once and never return to again (what is Chinook?).